A Geostatistical Evolution Strategy for Subsurface Characterization: Theory and Validation Through Hypothetical Two‐Dimensional Hydraulic Conductivity Fields
Author:
Open Access
- 26 February 2020
- journal article
- research article
- Published by American Geophysical Union (AGU) in Water Resources Research
- Vol. 56 (3)
Abstract
A novel approach that uses the covariance matrix adaptation evolution strategy in conjunction with geostatistical covariance functions and simultaneous ensemble Kalman filters is introduced to characterize the hydraulic conductivity fields of aquifers. To verify the performance, several similarity measures between the target and estimated fields are used, including the correlation coefficient, least absolute deviation, least squares, and cross‐variance ratio. Three different cases are synthesized through validation with a small‐scale domain (40 × 40), in which the hydraulic conductivity fields are stationary, nonstationary, and stationary with reduced prior information. According to these applications, the developed approach produces excellent estimations whether the field is stationary or nonstationary when the conditioning information is sufficient. With limited conditioning information, the performance of the estimation is slightly but not significantly lower. To validate the scalability, the developed approach is applied to larger domains (100 × 100, 150 × 150, and 200 × 200), and a modified approach based on principal components is used. According to this analysis, the developed approach requires low computational cost by making predictions with a limited number of forward‐model runs and internal computations. Furthermore, the approach requires a small amount of prior information for the estimation. Based on these applications, this approach shows high potential to create cost‐effective estimates that can be used as a reference for the current state‐of‐the‐art geostatistical inversion approaches to improve subsurface characterization.Keywords
Funding Information
- Korea Environmental Industry and Technology Institute (2018001810004)
This publication has 42 references indexed in Scilit:
- Bayesian approach for three-dimensional aquifer characterization at the Hanford 300 AreaHydrology and Earth System Sciences, 2010
- On Integrating Groundwater Transport Models with Wireless Sensor NetworksGroundwater, 2010
- Optimized groundwater drawdown in a subsiding urban mining areaJournal of Hydrology, 2009
- Geostatistical inference of hydraulic conductivity and dispersivities from hydraulic heads and tracer dataWater Resources Research, 2006
- A modified Levenberg–Marquardt algorithm for quasi-linear geostatistical inversingAdvances in Water Resources, 2004
- Reducing the Time Complexity of the Derandomized Evolution Strategy with Covariance Matrix Adaptation (CMA-ES)Evolutionary Computation, 2003
- Completely Derandomized Self-Adaptation in Evolution StrategiesEvolutionary Computation, 2001
- Bayesian Method for hydrogeological site characterization using borehole and geophysical survey data: Theory and application to the Lawrence Livermore National Laboratory Superfund SiteWater Resources Research, 1999
- An Iterative Stochastic Inverse Method: Conditional Effective Transmissivity and Hydraulic Head FieldsWater Resources Research, 1996
- A geostatistical approach to the inverse problem in groundwater modeling (steady state) and one‐dimensional simulationsWater Resources Research, 1983